Existing package splits into dependent subpackages -- best packaging practice ?
Hi all,
I have long maintained rpy2 (and rpy before) which provides a bridge from
Python to R (which I tend to care more for), and am on friendly terms with
its author. You can see my repo at salsa [1], it is pretty vanilla.
The upcoming upstream release will split into three packages, all in the same
source repo [2]. I am a casual Python user, and not all into packaging there
(source or for Debian). But I can wondering that this arrangement must exist
elsewhere. Is there a good pattern I can borrow to build (and then install ?)
rpy2-interfaces to then build (and install ?) rpy2-objects to then build
rpy2?
Python users see these as independent as they are in three different PyPI
packages.
Is there a best or recommended way to approach this? CCs welcome, I am not
subscribed to debian-python.
Cheers, Dirk
[1] https://salsa.debian.org/edd/rpy2
[2] https://github.com/rpy2/rpy2
--
dirk.eddelbuettel.com | @eddelbuettel | edd@debian.org
Reply to: